Custom speaker enclosure, as innert as possible

[DanielT]

Bracing or not that is the question. There are a number of parameters/factors to consider such as building materials, thickness of the walls and the size and shape of the boxes.There are probably more factors, that's just what I came up with now.

Take size and bracing.
Bracing for these, most likely it is not needed:

Since Tannoy was mentioned in thread. Bracing for these? For all? Only for the biggest? Hopefully Tannoy tested this by building different test boxes with and without bracing and different types of bracing:

 

[MAB]

I'm just going to say that in my time on DIYAudio.com, I've seen a fair number of before / after measurements of bracing and it tends to be pretty visible in measurements. Anecdotally a lot of people claim the changes are audible. And it's very uncontroversial in the speaker-building world that bracing is necessary to prevent unwanted resonances.

I don't think I've ever seen blind tests of braced vs. unbraced speaker cabinets... but at the end of the day the measured differences are often big enough to be clearly audible in theory. I don't think it's silly to do that kind of work on cabinets "just in case", certainly more sensible than upgrading from a 90 to 110db SINAD DAC.

Yeah, I used to be a member there. I saw lots of testimonials. Occasionally measurements, although often lacking contols with setup changes between runs. The few that had large response changes were where they also changed port length, or they confused a change in box tuning due to the decrease in volume, o had the mic in wildly different placements.

I have two identical boxes for a pair of 15" woofers. The boxes are made from 1/2" void-free birch ply. The boxes are prototypes. One is unbraced except for a circular cutout glued to the back wall, as shown below:

The other cabinet is identical, except each corner is braced with 2" chamfered braces all around, filled with polyurethane epoxy, with jute glued on. Like this:

I used polyurethane glue on all of the jute, so I can't rip it out. The cabinets are 20" x 20" x 12".

I am finishing the final versions, made from 1" Baltic birch. They will be the same internal volume, braced. The braces are as much to get the whole cabinet lined up as I glue them. I'll measure them later. And provide more detailed interior phots. For now, just the two prototype cabinets.

I put a block of wood approximately equal to the volume of bracing materials into the unbraced cabinet to try to tune them to the same response.

I made 4 sets of measurements. I used the same woofer throughout, using an electric driver to get consistent torque, set to eleven of course .
Experiment a: Woofer in the unbraced cabinet, no stuffing, absolutely empty
Experiment b: Woofer in the braced cabinet with the glued on jute mat but no stuffing
Experiment c: Woofer in the bare cabinet with jute mat on the back wall only and stuffing
Experiment d: Woofer in the braced cabinet with the glued on jute mat and stuffing

Here are experiment a and b. It looks like I got the size of the blosck right, since the resonance peak at 70 Hz is the same. I am guessing the 5 Ohm lower peak impedance at 70Hz is the jute mat in the braced cabinet. There are some small variations in the impedance trace of the empty unbraced cabinet.

Including experiment c and d, where I add a roughly equivalent amount of stuffing to each cabinet reduces the box tuning to about 67Hz.

Zooming in:

The unbraced box (expt b) has resonances at 280Hz, 360Hz, 640Hz, 740Hz, and 860Hz. The braced box eliminates the resonances at 280Hz and 360Hz, but those resonances remain in the unbraced box when I stuff it. The resonances at 640, 740, and 860Hz are all controlled in experiment b-d. From the impedance traces, it looks like the resonances at 240 and 360Hz are panel resonances since they go away in the braced box. The three higher frequency resonances all are smoothed with stuffing or jute.

How audible is this? I recorded each configuration nearfield. I was fairly careful to get the mic distance the same each time.

The resonances at 640, 740, and 860Hz all show up in the frequency response of the empty box (expt b). The resonance at 240Hz doesn't show up in the frequency response, a peak at 340Hz shows up in all of the frequency response no matter braced or not, and has nothing to do with the 360Hz impedance resonance. This 340Hz peak is the cone surround resonance frequency, and isn't affected by the cabinet bracing or stuffing.

This makes a case for stuffing a bare box, that's where the biggest change is.

Lastly the distortion from the four runs:

Some tiny and totally nitpicking differences. Even the unstuffed cabinet is low distortion.

The unbraced speaker doesn't perform much different than the braced one. The only experiment I did here that is potentially audible is the unstuffed and unbraced cabinet up around 600-700Hz.

When I get the final cabinets finished I will post measurements of those as well. I don't think they will be much different. And while I hope I am happy with the finals, I just don't think the premium 1" Baltic birch is going to change the sound. But I am hoping they will have a nice finish.

 

[DanielT]

Speaking of sand since it is mentioned in the thread. I wonder how two cardboard casting tubes with different diameters would work. If you put the smaller tube in the larger one and fill the space between them with sand or concrete. The sound will bounce in a tube, no standing waves. Considering all the sand or concrete between the pipes, it should be really stable speakers.It would maybe work well? The pipes themselves are not expensive. The challenge, I think, then lies mostly in getting the baffle with the speaker elements mounted on the tube or tubes in a sensible and stable way.
It might be quite heavy speakers, but with sand you can mount drain plugs for the sand. Pour in and fill out if necessary.

Just a thought.

 

[Mr. Widget]

...and moderate quantities of (very unsexy) polyfill are what is needed.

A bit off topic, but possibly useful in DIY projects...

Years (decades really) before I had proper test equipment I discovered that acoustic foam, acoustic fiberglass, and Ultratouch (recycled denim house insulation) were all substantially better than Dacron batting at mitigating standing waves and other unwanted cabinet issues. I have no measurements to support this, but back in the day I compared many materials and these were blind tests in that I didn't know what was in the various cabinets, and the polyfill was consistently less effective than these other materials.

If someone is working on a DIY project they could easily compare and measure the differences, I am confident they will be quite obvious if properly measured.

 

[MAB]

A bit off topic, but possibly useful in DIY projects...

Years (decades really) before I had proper test equipment I discovered that acoustic foam, acoustic fiberglass, and Ultratouch (recycled denim house insulation) were all substantially better than Dacron batting at mitigating standing waves and other unwanted cabinet issues. I have no measurements to support this, but back in the day I compared many materials and these were blind tests in that I didn't know what was in the various cabinets, and the polyfill was consistently less effective than these other materials.

If someone is working on a DIY project they could easily compare and measure the differences, I am confident they will be quite obvious if properly measured.

Thanks. I might measure some of these options. Agreed, some materials are better than others for standing waves. Interestingly, jute mat is not that effective at standing wavers, not sure why I am still using it except I still have a few large rolls. I thought there was a thread by @Rick Sykora where lots of different materials and fill quantities were measured. I seem to recall the focus was on changing the box tuning. I can't find it for the life of me...

 

[Rick Sykora]

Likely, as some of my damping experiments are buried in Directiva threads, perhaps this is the one…

Introducing Directiva - An ASR open source platform speaker project

Here is the nearfield on the right PR with different levels of damping... First, is the new baseline when I pulled the Sonic Barrier from the cabinet (red trace). While you can see some small differences in spots, note the peaks between 2-4 kHz. These are significantly reduced with addition...

www.audiosciencereview.com

I did a more controlled analysis of vent modifications in Understanding Vented Speakers thread. There are some interesting youtuber experiments posted too. Available types of damping vary regionally as we found for Directiva r2 but its alignment is a very special case.

 

[kemmler3D]

Yeah, I used to be a member there. I saw lots of testimonials. Occasionally measurements, although often lacking contols with setup changes between runs. The few that had large response changes were where they also changed port length, or they confused a change in box tuning due to the decrease in volume, o had the mic in wildly different placements.

I have two identical boxes for a pair of 15" woofers. The boxes are made from 1/2" void-free birch ply. The boxes are prototypes. One is unbraced except for a circular cutout glued to the back wall, as shown below:
View attachment 387111

The other cabinet is identical, except each corner is braced with 2" chamfered braces all around, filled with polyurethane epoxy, with jute glued on. Like this:
View attachment 387117
I used polyurethane glue on all of the jute, so I can't rip it out. The cabinets are 20" x 20" x 12".

I am finishing the final versions, made from 1" Baltic birch. They will be the same internal volume, braced. The braces are as much to get the whole cabinet lined up as I glue them. I'll measure them later. And provide more detailed interior phots. For now, just the two prototype cabinets.

I put a block of wood approximately equal to the volume of bracing materials into the unbraced cabinet to try to tune them to the same response.

I made 4 sets of measurements. I used the same woofer throughout, using an electric driver to get consistent torque, set to eleven of course .
Experiment a: Woofer in the unbraced cabinet, no stuffing, absolutely empty
Experiment b: Woofer in the braced cabinet with the glued on jute mat but no stuffing
Experiment c: Woofer in the bare cabinet with jute mat on the back wall only and stuffing
Experiment d: Woofer in the braced cabinet with the glued on jute mat and stuffing

Here are experiment a and b. It looks like I got the size of the blosck right, since the resonance peak at 70 Hz is the same. I am guessing the 5 Ohm lower peak impedance at 70Hz is the jute mat in the braced cabinet. There are some small variations in the impedance trace of the empty unbraced cabinet.


View attachment 387120
Including experiment c and d, where I add a roughly equivalent amount of stuffing to each cabinet reduces the box tuning to about 67Hz.

View attachment 387122
Zooming in:
View attachment 387123
The unbraced box (expt b) has resonances at 280Hz, 360Hz, 640Hz, 740Hz, and 860Hz. The braced box eliminates the resonances at 280Hz and 360Hz, but those resonances remain in the unbraced box when I stuff it. The resonances at 640, 740, and 860Hz are all controlled in experiment b-d. From the impedance traces, it looks like the resonances at 240 and 360Hz are panel resonances since they go away in the braced box. The three higher frequency resonances all are smoothed with stuffing or jute.

How audible is this? I recorded each configuration nearfield. I was fairly careful to get the mic distance the same each time.
View attachment 387125

The resonances at 640, 740, and 860Hz all show up in the frequency response of the empty box (expt b). The resonance at 240Hz doesn't show up in the frequency response, a peak at 340Hz shows up in all of the frequency response no matter braced or not, and has nothing to do with the 360Hz impedance resonance. This 340Hz peak is the cone surround resonance frequency, and isn't affected by the cabinet bracing or stuffing.

This makes a case for stuffing a bare box, that's where the biggest change is.

Lastly the distortion from the four runs:
View attachment 387127
Some tiny and totally nitpicking differences. Even the unstuffed cabinet is low distortion.

The unbraced speaker doesn't perform much different than the braced one. The only experiment I did here that is potentially audible is the unstuffed and unbraced cabinet up around 600-700Hz.

When I get the final cabinets finished I will post measurements of those as well. I don't think they will be much different. And while I hope I am happy with the finals, I just don't think the premium 1" Baltic birch is going to change the sound. But I am hoping they will have a nice finish.

This is great info, I would just say that I think if anyone's having problems with audible cabinet resonances it tends to be in the low-mids and it would probably be using less sturdy walls than you've used here, e.g. 16mm MDF or worse. 1" BB is fairly "high end" in terms of wall materials AFAIK.

 

[Salt]

It would need to explore every single square cm with a close to microphone or accelerator to examine a difference ... that will be there, but who will work it out?

 

[MAB]

This is great info, I would just say that I think if anyone's having problems with audible cabinet resonances it tends to be in the low-mids and it would probably be using less sturdy walls than you've used here, e.g. 16mm MDF or worse. 1" BB is fairly "high end" in terms of wall materials AFAIK.

These cabinets are 1/2" voidless 5-ply birch. I would call them mid-grade in the birch-ply scheme of things. The woofer boxes are small. This is my upper limit for box-size using 1/2", but that is my gut feeling.

The 1" Baltic birch is super-fine. I am mostly getting it for the finish, but it makes for a satisfyingly strong and dead speaker. And it is great to work with. But I am eager to see if it is fundamentally different or better than the lighter boxes. I think overbuilt boxes are great, but not just for the sake of overbuilding. Methods or materials that make it harder to build and finish can be a drag.

 

[Dumu]

Interesting topic!

I have used carpenter's clamps to squeeze thick boards wrapped in towels to the sides of cabinets to see if that would cure suspected resonances. Has been working quite well for me.

I own a pair of klipsch floorstanders with 2x 8" woofers that have a really huge sound, but that had nasty echo/boomy bass in my room that I was able to mostly cure by lining the cabinets with self adhesive sound deadening and adding loose polyfill. Cabinets were actually pretty well braced from factory, they just were empty. The speakers lost some of that huge soundfield, but I still think it was worth it. I also learned that it is possible to add too much polyfill (soundfield became super narrow and it was like listening to pair of heaphones). Knock test went from really hollow-sounding to almost inert.